MPSO-Type-3/main.py

"""
MIT License

Copyright (c) 2020 Bachtiar Herdianto

Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all
copies or substantial portions of the Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
SOFTWARE.
"""
import numpy as np
import random
import matplotlib.pyplot as plt
import time

def Sphere (x): # Sphere function
    total = 0
    for i in range(len(x)):
        total += x[i]**2
    return total

def Rastrigin (x):  # Rastrigin function
    total = 10*len(x)
    for i in range(len(x)):
        total += x[i]**2 - (10*np.cos(2*np.pi*x[i]))
    return total

def grad_Sphere (x):    # Derivative of Sphere function
    gradient_coordinate = []
    for i in range(len(x)):
        total = 2*x[i]
        gradient_coordinate.append(total)
    return np.array(gradient_coordinate)

def grad_Rastrigin (x):     # Derivative of Rastrigin function
    gradient_coordinate = []
    for i in range(len(x)):
        total = 2*x[i] + 10*2*np.pi*x[i]*np.sin(2*np.pi*x[i])
        gradient_coordinate.append(total)
    return np.array(gradient_coordinate)

class Particle:
    def __init__(self, dim, minx, maxx, error):
        self.position = np.random.uniform(low=minx, high=maxx, size=dim)
        self.velocity = np.random.uniform(low=minx, high=maxx, size=dim)
        self.best_part_pos = self.position.copy()
        self.error = error(self.position)
        self.best_part_err = self.error.copy()
    def setPos(self, pos, error):
        self.position = pos
        self.error = error(pos)
        if self.error < self.best_part_err:
            self.best_part_err = self.error
            self.best_part_pos = pos
    def controlPos(self, bounds):
        for i in range(len(bounds)):
            if self.position[i] < bounds[i][0]:
                self.position[i] = bounds[i][0]
            if self.position[i] > bounds[i][1]:
                self.position[i] = bounds[i][1]

class PSO:
    def __init__(self, dims, numOfParticles, numOfEpochs, lower, upper, funct, grad):
        self.swarm_list = [Particle(dims, lower, upper, funct) for i in range(numOfParticles)]
        self.numOfEpochs = numOfEpochs
        self.best_swarm_position = np.random.uniform(low=lower, high=upper, size=dims)
        self.dimension = dims
        self.upper = upper
        self.lower = lower
        self.best_swarm_error = -1
        self.function = funct
        self.gradien = grad
        self.boundaries = []
        for i in range(dims):
            self.boundaries.append((lower, upper))

""" Optimize Function """
def optimize_signature(self, w, c1, c2):
    r1 = random.random()
    r2 = random.random()
    funct = self.function
    X_epoch = []
    Y_error = []
    boundaries = self.boundaries
    for i in range(self.numOfEpochs):
        for j in range(len(self.swarm_list)):
            current_particle = self.swarm_list[j]
            Vcog = r1*c1*(current_particle.best_part_pos - current_particle.position)
            Vsos = r2*c2*(self.best_swarm_position - current_particle.position)
            deltaV = w*current_particle.velocity + Vcog + Vsos   # calculate deltaV
            new_position = current_particle.position + deltaV # calculate the new position
            self.swarm_list[j].setPos(new_position, funct)
            self.swarm_list[j].velocity = deltaV
            self.swarm_list[j].controlPos(boundaries)

            # check the position if it is best for swarm
            if funct(new_position) < self.best_swarm_error or self.best_swarm_error == -1:
                self.best_swarm_position = new_position
                self.best_swarm_error = funct(new_position)
        X_epoch.append(i)
        Y_error.append(self.best_swarm_error)

    a, b = np.array(X_epoch), np.array(Y_error)
    plt.plot(a, b)
    plt.xlabel('Epoch'), plt.ylabel('Objective value (Error)'), plt.title('Report Optimization'), plt.show()
    print('----------------------------\nTotal Epoch: {0} \nBest position: \n[{1}, {2}, {3}, {4}, {5}, {6}] \nBest known error: {7}'.format(
        i + 1, self.best_swarm_position[0], self.best_swarm_position[1], self.best_swarm_position[2],
        self.best_swarm_position[3], self.best_swarm_position[4], self.best_swarm_position[5], self.best_swarm_error))

def optimize_grad(self, w, c1, c2, lr):
    funct = self.function
    grad_funct = self.gradien
    boundaries = self.boundaries
    X_epoch = []
    Y_error = []
    for i in range(self.numOfEpochs):
        for j in range(len(self.swarm_list)):
            current_particle = self.swarm_list[j]   # get current particle
            Vcurr = grad_funct(current_particle.position) # calculate current velocity of particle
            Vcog = c1*(current_particle.best_part_pos - current_particle.position)
            Vsos = c2*(self.best_swarm_position - current_particle.position)
            deltaV = w*Vcurr + Vcog + Vsos   # calculate deltaV
            new_position = self.swarm_list[j].position - lr*deltaV # calculate the new position
            self.swarm_list[j].setPos(new_position, funct)     # update the position of particle
            self.swarm_list[j].controlPos(boundaries)

            # check the position if it is best for swarm
            if funct(new_position) < self.best_swarm_error or self.best_swarm_error == -1:
                self.best_swarm_position = new_position
                self.best_swarm_error = funct(new_position)
        X_epoch.append(i)
        Y_error.append(self.best_swarm_error)

    a, b = np.array(X_epoch), np.array(Y_error)
    plt.plot(a, b)
    plt.xlabel('Epoch'), plt.ylabel('Objective value (Error)'), plt.title('Report Optimization'), plt.show()
    print('----------------------------\nTotal Epoch: {0} \nBest position: \n[{1}, {2}, {3}, {4}, {5}, {6}] \nBest known error: {7}'.format(
        i + 1, self.best_swarm_position[0], self.best_swarm_position[1], self.best_swarm_position[2],
        self.best_swarm_position[3], self.best_swarm_position[4], self.best_swarm_position[5], self.best_swarm_error))

""" Running Code 
    General setting:
    Number of dimmensions: 6
    Number of particles: 30
    Number of Max Iterations: 500
    
    Setting no. 1:
    Objective Funct.: Rastrigin
    
    Setting no. 2:
    Objective Funct.: Sphere  """
setting01 = PSO(
    dims=6, 
    numOfParticles=30, 
    numOfEpochs=500, 
    lower=-500, 
    upper=500, 
    funct=Rastrigin, 
    grad=grad_Rastrigin
    )

setting02 = PSO(
    dims=6, 
    numOfParticles=30, 
    numOfEpochs=500, 
    lower=-500, 
    upper=500, 
    funct=Sphere, 
    grad=grad_Sphere
    )

def main():
    print('Original PSO using setting no. 1:')
    start1 = time.time()
    optimize_signature(setting01, w=0.729, c1=1.49445, c2=1.49445)
    end1 = time.time()
    print("processing time:", end1-start1, 'second\n---------------------\n')

    print('Original PSO using setting no. 2:')
    start2 = time.time()
    optimize_signature(setting02, w=1, c1=2, c2=2)
    end2 = time.time()
    print("processing time:", end2-start2, 'second\n---------------------\n')

    print('Modified Weight-Addictive PSO:')
    start = time.time()
    optimize_grad(setting01, w=0.729, c1=1.49445, c2=1.49445, lr=0.01)
    end = time.time()
    print("processing time:", end - start, 'second\n')


main()